Hu, Stephane W., Marchal, Guilhem, Dilem, Yvan, Walch, Denis et Tabiai, Ilyass.
2026.
« Planar validation of speckle pattern optimization for large-scale 3D streamlined structures in stereo digital image correlation using anamorphic transformation ».
Optics and Lasers in Engineering, vol. 202.
Prévisualisation |
PDF
Tabiai-I-2026-33616.pdf - Version publiée Licence d'utilisation : Creative Commons CC BY. Télécharger (10MB) | Prévisualisation |
Résumé
This study addresses limitations of stereo Digital Image Correlation (DIC) when applied to large objects viewed from tilted angles (LOVTA), such as aircraft wings and fuselages. Under such configurations, perspective distortions lead to strong variations in apparent speckle size and spacing, which degrade correlation robustness and increase measurement uncertainty when conventional speckle patterns are used. To mitigate these effects, a geometric model is developed to generate anamorphic speckle patterns that explicitly pre-compensate for the known imaging geometry of the DIC setup. By locally transforming the reference pattern according to camera orientation and surface geometry, the proposed approach aims to maintain more uniform speckle characteristics in the image plane, thereby improving correlation conditioning. An experimental comparison between regular and anamorphic speckle patterns is conducted on a planar target subjected to rigid-body rotations, under grazing-angle observation conditions representative of full-scale aircraft testing. The DIC analysis is used to extract the 3D positions of subset centers, from which uncertainty metrics are derived to assess correlation precision rather than to measure actual displacement or strain fields. The results show that the anamorphic speckle pattern consistently reduces matching uncertainty and improves the precision of 3D point localization compared to a regular pattern, even as camera incidence angles increase. The study intentionally focuses on a controlled planar configuration to isolate perspective-induced effects and establish a clear baseline for validation. Although the transformation framework is formulated generically and illustrated analytically for curved surfaces, experimental validation of complex geometries and full-scale structures is identified as future work. The anamorphic transformation is applied offline during pattern generation and does not affect the DIC correlation workflow. Despite practical challenges related to pattern generation, application, and geometric modeling accuracy, the results demonstrate that geometric pre-compensation at the speckle-design stage is a promising and physically grounded strategy for reducing uncertainty in DIC measurements of large structures observed under strong perspective conditions.
| Type de document: | Article publié dans une revue, révisé par les pairs |
|---|---|
| Chercheur(-euse): | Chercheur(-euse) Tabiai, Ilyass |
| Affiliation: | Génie mécanique |
| Date de dépôt: | 17 avr. 2026 20:37 |
| Dernière modification: | 22 avr. 2026 20:47 |
| URI: | https://espace2.etsmtl.ca/id/eprint/33616 |
Actions (Authentification requise)
 |
Dernière vérification avant le dépôt |